2. • “Digital consumption has generated a huge variety of new opportunities for
marketers—but insights and intelligence are not growing as much as data points,
as a culture of short termism prevails. We recommend the linking of audience
measurement and consumer behaviour data, but the industry lacks both
standards and trust, while the still-immature digital marketing supply chain
poses problems for data integrity.”
Enders Analysis – Marketing and Measurement in the Digital Era, 2016
Time to measure audience attitudes and track their behaviour
3. The journey to the digital holy land in numbers: A tough pilgrimage for
everyone apart from Facebook and Google
2011
Digital ad spend
in the UK
overtakes TV
£4.8bn
2013
Newspaper websites
reach 70% of online
adults
33 mill.
uniques
2016
VC funding of ad
tech down.
Publishers see 70%
of CPM go to ad tech
-17%
deals
2017
Facebook and
Google take majority
of ad spend share…
and nearly all the
growth!
53%
share
2015
Programmatic =
61% of UK display
£1.85bn
4. FB and Google:
£4,987m
News digital display
£296m
Rest of digital pure play
display (and ad tech):
£2.942m
To put it in perspective…
5. How have we managed to get here? …. A perpetual preoccupation with The Click!
Measuring what we can not
what we should
Measuring short term rather
than long term impact
Measuring brand with
response metrics
6. This is where clicks get us: an industry in crisis?
of web traffic
is human
are blocking
adverts
of impressions are
viewable
44% 22% 40% 70%
1) AD FRAUD 2) AD BLOCKING 3) VIEWABILITY 4) VALUE OF PREMIUM
of premium publishers
CPMs lost to
programmatic
7. Even the big guys aren’t immune when it comes to measurement challenges
8. The key priorities for 2017…
74%
Very important to
growth
Strong capability
14%
For marketers:
Matching customers
across multiple
devices
61%
Are prioritising
data
For publishers:
Improving data
capabilities
60%
Of mobile ads
hitting their
target
For brands:
Verification
9. RAM Inviso provides a cost effective answer
• Panel based verificationFRAUD
• Measure which ads engageAD BLOCKING
• Who’s actually paying attention to your ads?VIEWABILITY
• Establish benchmarks and optimiseVALUE OF PREMIUM
• True de-duplicated reachCROSS PLATFORM
THE
ISSUE
INVISO
12. Case studies from three countries and six categories
AUTOMOTIVE
B2B FINANCE RETAIL
FASHIONTRAVEL
13. 1) We’ve all seen a campaign delivery report…. But de-duplicated across devices
and across users?
B2B Campaign
Total Reach:
397,000
Fashion
Campaign Total
Reach:
114,000
14. Measuring true campaign delivery can be cost effective and scalable, enabling us
to draw insight from a whole host of different campaign profiles
19. 2) Verification is key. Especially when targeting still leaves a lot to be desired
MANAGEMENT
21%
HIGH INCOME
8%
CAR OWNER
90%
KIDS IN HOUSEHOLD
27%
AGED 16-34
15%
PURCHASE INFUENCER
84%
% Profile
20. When it comes to mainstream shopper audiences, mobile seems to be hitting the
mark
103
103
98
Mobile
Tablet
Desktop
139
108
98
Mobile
Tablet
Desktop
113
87
90
Mobile
Tablet
Desktop
111
98
Tablet
Desktop
11
64
119
Mobile
Tablet
Desktop
118
98
Tablet
Desktop
MANAGEMENT HIGH INCOME CAR OWNER KIDS IN HOUSEHOLD AGED 16-34 PURCHASE INFUENCER
Profile index
21. 3) Brand impact measurement can be scalable and cost effective… and
benchmarkable
22. (Just in case there was any doubt as to why brand measurement matters…)
Recommended Long
Term vs Short Term
ad spend for brands:
60:40
Estimated brand volume
resulting from short
term metrics:
5%
Source: Field and Binet analysis of IPA databank. Millward Brown
23. 4) As they’re based on observed rather than self reported media usage data,
Inviso tests tend to result in a higher ad recall
Standard
35%
Inviso
38%
24. 4) Analytics and surveyed data become greater than the sum of their parts:
frequency gives time for new information to bed in?
57%
Discovered new
information
Benchmark
39%
26. An evenly delivered campaign avoids irritation through over exposure?
50%
Positive
towards ad
Benchmark
38%
27. A phased campaign with offline synergies boosts brand recognition
69%
Easily identified
advertiser
Benchmark
51%
28. What we know now…
• We can tackle the issues that pervade the digital industry through enhanced measurement
capabilities
• Campaign reach can be de-duplicated across devices more easily than ever before
• Campaign verification is essential. Most campaigns still don’t hit their target
• Mobile does offer a world of new targeting capabilities to mainstream audiences however
• By linking behaviours we have insight in to a whole host of campaign impact behaviours
• Behaviours and attitudes combined are greater than the sum of their parts!
29. Thank you!
If you’d like a copy of this report please email
dianne.newman@rampanel.com
Notas del editor
- We live in a world of data automation. An algorithmically ruled world in which the reputation of data is increasingly playing second fiddle to its automation.
As Cathy O Neil tells us in her book Weapons on Math Destruction, these algorithms – whether they decide our insurance premiums or dictate which ads we see – leave a lot to be desired when left unchecked.
As one of Europe’s largest suppliers of digital ad effectiveness research, it’s obviously upon advertising that we want to focus today. We’re going to explore what happens when the world of behavioural web data (or big data) is combined with more traditional survey based research to provide insight and transparency in to an advertising world which unfortunately increasingly has too little of either.
But first, a bit of context…..
We don’t feel we could put it much better than Enders do here really. There’s a big difference between data and insight in the digital world and one does not always lead to the other. The lack of transparency around ad data – much of which is based on the opaqueness of programmatic algorithms has lead to declining levels of trust and insight.
How do deal with this? By measuring audiences attitudes towards advertising and brands, and combining it with sophisticated behavioural web analytics data.
But let’s take a step back a second. How did we get to this point?
The digital holy land has been one to which any publisher, media owner or platform gripped in the forces of digital disruption has been journeying in an effort to create viable monetization strategies to ensure their future.
Six years ago the picture looked promising in the UK. Digital ad spend had just overtaken TV opening a whole wealth of ad revenue opportunities for those who knew anything about how to package up their audience and sell advertising. (Source: AA and IAB 2011)
By 2013, news brands were enjoying record levels of traffic, reaching 70% of UK adults overall. The scale that they had achieved together online was to rival Facebook and Google, so surely the ad pounds would follow? Source: Comscore December 2013
And then came the rise in programmatic which by 2015 accounted for 61% of all UK display spend. Programmatic presented a fantastic opportunity for publishers to monetise every last piece of remnant inventory on their sites, but at the same time placed a considerable downward pressure on CPMs…
Just a year later the Guardian estimated that it lost up to 70% of the value of an ad sold through ad exchanges vs when it was sold directly. At the same time, ad tech businesses themselves now face a squeeze and in 2016 for the first time, the number of venture capitalists investing in new ad tech firms declined. (Source: https://www.ft.com/content/c4c358ca-c6af-11e6-8f29-9445cac8966f . The number of global adtech venture financing deals will touch 343 by year-end, a 17 per cent drop compared with 414 deals in 2015, according to data collated for the Financial Times by CBInsights.)
The duopoly of Facebook and Google now looms large, taking over half of all UK online ad spend. Scale, data, enhanced measurement and provable ad attention are the pillars on which they have built there success. (https://www.theguardian.com/media/2016/dec/15/google-facebook-uk-online-ad-revenue)
Clearly there is an imperative for publishes and ad tech alike to up their measurement, data and reporting capabilities if they resist the profound changes taking place in the market.
- And just to really hammer the point home, Facebook and Google are said to have made nearly £5bn last year, leaving the rest of the market trailing in their wake.
Source: AA and emarketer 2016: part actuals and part forecast. 2016 FULL YEAR FIGURES
So how did publishers, digital pure plays and ad tech manage to be left so far behind Facebook and Google?
The answer is simple – as an industry we’ve been preoccupied with measuring ad impact using clicks.
We do this because it’s easy – not because it’s the right thing to do.
And sure, a click might be a suitable metric for a DR campaign… but for a brand campaign? How can you gain any insight I to how consumers feels about an ad or brand based on what less than one in a thousand people do when they actually see an online ad?
And where does this pre-occupation with clicks leads?
Well because it’s easy to game a click, ad fraud is reaching untold levels where it is now estimated by statista that less than half of all web traffic is actually human (the rest being bots and web spiders).
Because marketers are obsessed with acquiring more and more clicks, more and more irritating and intrusive ads are being served… which has lead to the rise in ad blocking: something that over 20% of the UK online population now do according to the IAB (weighted towards young men – making a hard to reach target audience even harder to reach).
People aren’t even looking at many ads any more, with viewability rates differing according to site design, ad load an ad positioning. The IAB and Group M both have their own standards of viewability, but does such a behavioural metrics create more harm than good?
And finally, the obsession with clicks is impacting content. More and more content is being created with the aim of generating an ad impression, rather than fulfilling a consumer need. More and more inventory is being created and most of it is non premium – yet do the programmatic algorithms differentiate between this and an ad on the Guardian or FT? Combined with all the fake news out there, premium content is a having a tough time!
So if clicks are what’s ruining the online advertising ecosystem then what is the answer?
Well often it hasn’t been easy to get measurement right. Even Facebook and Twitter can get it wrong!
A glance at any one of a number of industry surveys sees this measurement challenge playing out as a key priority for this year.
Marketers want to be able to match customers across multiple devices. Too few measurement systems de-duplicate reach by device or browser, creating huge inefficiencies in campaign delivery. While three quarters think this is a key priority, only 14% think they have a strong capability in this area.
For publishers improving data capabilities is key – both to generate new monetization opportunities, but also to truly understand their audience and drive maximise value for advertisers.
And for brands? Verification is crucial. The old adage goes that half of all ad spend is wasted, we just don’t which half. Brands are feeling exasperated by fraud and a lack of independent verification of A) who’s actually seeing their ads and B) What impact it’s having on their brand beyond simple clicks
(Sources: https://econsultancy.com/blog/68676-10-important-stats-from-econsultancy-s-2016-research/. Survey of 184 marketers
https://www.themediabriefing.com/article/publishers-eye-new-strategies-for-challenging-year-ahead Reuters Digital News Report
http://www.wsj.com/articles/mobile-ad-targeting-is-improving-according-to-nielsen-1479826801)
At RAM we believe we have just the solution. We can employ a unique, cost effective and scalable technology to provide publishers, platforms and ad tech with the independent verification they need to place trust back at the centre of the online ad industry.
Using a panel based audience verification system we can tell you who’s seeing your ads and how effectively they’re reaching your target.
If people are blocking ads then why? We need to create ads that engage not irritate. Our scalable survey platform can do this for you.
What does viewability tell you about your ads anyway? Lumen suggest that not all viewable ads are actually viewed, so why not measure actual ad recall based on a sample who know have been exposed?
For publishers, what premium should you place on advertising, how do ads on your site perform vs industry benchmarks… can these findings bee fed in to the programmatic algorithms?
And finally our unique panels allow us to deduplicate campaign reach and impact across devices – giving a true sense of unique reach. We’ve got some really interesting case studies for you in this area to come.
- RAM INVISO METHODOLOGY IN MORE DETAIL.
Although normative databases are powerful tools in helping us understand how advertising works, they sometimes lack to the subtlety to unpick micro trends in campaign performance.
We’ve therefore chosen six randomly selected case studies from our database of Inviso tests unpick what happens at a granular level when viewing campaign performance from both a behavioural and attitudinal point of view.
The findings – particularly in the context of the challenges our industry faces – are really interesting.
- We’ve all seen an ad server generated delivery report like the one of the left before haven’t we?
Increasingly, reporting on impressions is not enough however. Only by de-duplicating reach on an individual basis according to the device they’ve been exposed on can we get a true sense of campaign reach: such as the 397k this b2b campaign reached or the 114k this fashion campaign reached.
The biggest challenge with this has been about producing cross-device deduplication reports in a cost effective scaleable way, but with Inviso panel based solution, theses problems are removed….
- The Inviso methodology is subtle enough to pick up campaign delivery profiles of all different shapes and sizes.
For example, the two day takeover of a major publisher’s homepage. As you’d expect – and as was presumably the point – the campaign hit the vast majority of its audience after just a single day and then topped out on the second day.
Standard frequency builder campaigns are picked up too. In this campaign 100% of the audience was reached after a couple of weeks, but it still ran for another seven days accumulating frequency of exposure. As we all know, there are many reasons why brands want to do this, but as we’ll see later we can definitively link this profile to brand impact now.
The behavioural economist and psychologist Daniel Kahnemann in his book How Brands Grow – tells us that market success is built on market penetration. And the first step to that? Awareness building – like this steadily delivered B2B campaign which is aiming for reach over frequency.
The video campaign has a “stop-start” profile, with a couple of delivery spikes coinciding with a national TV campaign. Cross platform synergies should never be overlooked!
So we know how many people have been reached…. But who are they? Sadly in four out of the six cases, not who the advertisers was actually paying for.
The two campaigns which did achieve more than 50% of their target were also aimed at quite, mainstream audiences.
Given the challenges discussed earlier, perhaps this is not surprising, especially when according to Nielsen , only 53% of the impressions served in the UK last year reached their intended target (http://www.nielsen.com/uk/en/press-room/2016/nearly-half-of-online-campaigns-miss-target-audience.html).
Although a small, but randomly selected sample, these results look even worse than that.
Big Data has shifted the dial when it comes to the automation of the digital ad economy, placing ads in new, weird and wonderful ways in an Nth of a second like never before. The weakest link in an algorithm however is invariably the human which created. Combined with the issue of of fraud it’s unsurprising so many campaigns still fall wide of the mark.
When looking at this picture by platform, some interesting trends emerge: For starters mobile is actually stacking up pretty well for more mainstream audiences. The smartphone is so engrained in our daily lives that it is a vital tool in the path to purchase and part of people’s shopping and product research habits whether they’re in or out of home. To find such key shopper audiences (i.e.. kids in household, 16-34 and purchase influencers) delivering well on mobile make a good deal of sense.
When it comes to a higher incomes audiences, as with our B2B campaign, this trend is less evident. Of course it’s not that higher incomes earners aren’t to found on mobile, but would you be better targeting them on desktop during their working day?
We’ve looked at behavioural data so far, and at panellists profile data for verification…. But where does attitudinal data come in to it?
This where Inviso really comes in to its own by allowing us to measure audience attitudes and track their behaviour
RAM have tested thousand of digital campaigns, building up extensive benchmarks for how digital advertising works. Because it employs a self-reported single cell methodology, making a comparison to a benchmark rather than between control and exposed groups, it’s far easier to build up benchmarks rapidly.
Here we can see for example, that the average recall of a online ad is 35%, prompting a purchase intent among 11% of respondents.
And just in case we needed reminding (and seemingly we do when our industry is so obsessed with short term measures): investing in brand (and therefore brand impact measurement) matters.
An analysis if the IPA databank has shown that those advertisers who’s invest 60% of their budget in brand and 40% in short term (like direct response) marketing achieve far greater business benefits than those who invest two much in short term success.
Millward Brown further proves this point by claiming that only 5% of brand volume sales can be attributed to short terms metrics.
Awareness, understanding, favourability, purchase intent…. Brand matters!
Because brand matters, measuring it is so important to get right.
The standard method of measuring ad recall takes those who claim to have visited a web page on a specific data range and asks about their recall accordingly. This method relies of the imperfections of the human memory of course.
With Inviso however… we know when and where they have seen the campaign… subsequently giving us a more accurate base on which to measure recall.
Not only is this more accurate, but we’re also seeing a marginal improvement in ad recall rates when using Inviso measurement.
When it comes to brand impact, cross referencing it against the profile of the campaign delivered can unearth rich new insights that many marketers still just don’t have at their disposal.
From the campaigns we’ve looked at, there is a sense that planning instinct is still correct, although we can’t definitely say how much impact is down to the creative vs the campaign delivery, it’s interesting to note that our frequency builder campaign has outperformed the established benchmarks in consumers discovering new information about the brand. Frequency of exposure is used to bed in deeper brand experiences and we can see that being played out here.
(benchmarked against similar category campaigns)
From a simple recall level, the homepage takeover has proven to be highly memorable. – more than double the benchmark.
- A campaign which doesn’t overexpose itself on the other hand is less likely to irritate consumers and we’ve seen it work harder when it comes to brand persuasion.
- In the case of the retail video ad, synergies with an offline campaign can make brand identification that much easier.
Benchmarked against other online campaigns from last two years from same country
Benchmarked against other online campaigns from last two years from same country